Struct Tryll::Client::VoiceInputConfig

ClassList > Tryll > Client > VoiceInputConfig

Parameters for creating a VoiceInput handle.

• #include <VoiceInput.h>

Public Attributes

Type	Name
float	hotwordsScore = 1.5f
std::string	hotwordsStoragePath
AudioFormat	inputFormat Format of samples passed to SendAudioBuffer.
std::string	modelName Catalog name, e.g. "Whisper Tiny EN (int8)".
std::uint32_t	vadMinSilenceMs = 500 Silence duration (ms) that closes a segment.
std::uint32_t	vadSpeechPadMs = 250
float	vadThreshold = 0.5f Silero VAD speech-probability threshold.

Public Attributes Documentation

variable hotwordsScore

float Tryll::Client::VoiceInputConfig::hotwordsScore;

Bonus score added to each matched hotword token during decoding. Typical range 1.0–2.5. Ignored when hotwordsStoragePath is empty.

---

variable hotwordsStoragePath

std::string Tryll::Client::VoiceInputConfig::hotwordsStoragePath;

Relative path (resolved against the session storage folder) to a List StringStorage (.txt) whose entries are hotword phrases that the recognizer should be biased toward. Empty = no hotword biasing. The file loads on demand server-side — no prior CreateStringStorage needed.

---

variable inputFormat

Format of samples passed to SendAudioBuffer.

AudioFormat Tryll::Client::VoiceInputConfig::inputFormat;

---

variable modelName

Catalog name, e.g. "Whisper Tiny EN (int8)".

std::string Tryll::Client::VoiceInputConfig::modelName;

---

variable vadMinSilenceMs

Silence duration (ms) that closes a segment.

std::uint32_t Tryll::Client::VoiceInputConfig::vadMinSilenceMs;

---

variable vadSpeechPadMs

std::uint32_t Tryll::Client::VoiceInputConfig::vadSpeechPadMs;

Padding (ms) added around detected speech.

---

variable vadThreshold

Silero VAD speech-probability threshold.

float Tryll::Client::VoiceInputConfig::vadThreshold;

---

---

The documentation for this class was generated from the following file C:/_tryll/_monorepo3/tryll/clients/cpp/include/tryll/VoiceInput.h